AI-Driven Real-Time Heart Health Monitoring Using IoMT and LLM-Based Analytics

Authors:

Sadiya A, Syed Irfan A, Arram B, Daniya Y, Ashfan B, Pranay P

Abstract— Continuous monitoring of cardiovascular health plays a critical role in the early detection and prevention of cardiac conditions by enabling timely diagnosis. However, existing wearable systems often lack seamless integration with personalized, human-centric AI-based interpretation of the collected data. This research presents a comprehensive AI-driven platform that combines Internet of Things (IoT) technologies with advanced Large Language Models (LLMs) for real-time heart health monitoring. The system employs an AD8232 sensor integrated with an ESP32 microcontroller to acquire single-lead ECG signals at a sampling rate of 360 Hz. The acquired signals are digitized and transmitted via Bluetooth Low Energy (BLE) in optimized 12-sample packets to a custom Flutter-based mobile application, enabling real-time visualization and temporary local storage.

Subsequently, the data is synchronized with a cloud-hosted PostgreSQL database, where an automated background processing module performs physiological feature extraction, including R-peak detection, heart rate variability (HRV), and other waveform characteristics. These features are then analysed by a Model Context Protocol (MCP) server using a dual-prompt LLM framework. The system generates a quantitative heart health score from 0 to 10, along with an interpretable report for the user. The proposed architecture maintains low latency (less than 33 milliseconds per data packet), enabling efficient, reliable continuous monitoring. Overall, the system offers a scalable, innovative approach to proactive, personalized cardiovascular health management.

Keywords—Wearable IoT Devices, Real-Time Health Monitoring, Electrocardiogram (ECG) Analysis, Large Language Models (LLMs), Heart Rate Variability (HRV), Predictive Health Analytics